A viral protein disrupts vacuolar acidification to facilitate virus infection in plants

Abstract

Vacuolar acidification is essential for vacuoles in diverse physiological functions. However, its role in plant defense, and whether and how pathogens affect vacuolar acidification to promote infection remain unknown. Here, we show that Barley stripe mosaic virus (BSMV) replicase γa, but not its mutant γa_R569A, directly blocks acidification of vacuolar lumen and suppresses autophagic degradation to promote viral infection in plants. These were achieved via molecular interaction between γa and V‐ATPase catalytic subunit B2 (VHA‐B2), leading to disruption of the interaction between VHA‐B2 and V‐ATPase catalytic subunit E (VHA‐E), which impairs the membrane localization of VHA‐B2 and suppresses V‐ATPase activity. Furthermore, a mutant virus BSMV_R569A with the R569A point mutation possesses less viral pathogenicity. Interestingly, multiple viral infections block vacuolar acidification. These findings reveal that functional vacuolar acidification is required for plant antiviral defense and disruption of vacuolar acidification could be a general viral counter‐defense strategy employed by multiple viruses.

A viral protein disrupts vacuolar acidification to facilitate virus infection in plants.

Introduction

Vacuoles are the largest compartments in plant cells and possess various functions in plant growth, development, and response to abiotic and biotic stresses. All vegetative cells have lytic vacuoles containing hydrolytic enzymes and defense proteins, which suggests a potential role of vacuoles in plant defense. Plants have evolved at least two different systems, the vacuolar membrane‐collapse system and the vacuole–plasma membrane fusion system, to exploit vacuoles and vacuolar contents in order to defend disparate pathogens (Shimada et al, 2018). For instance, both systems are found to be associated with vacuolar membrane dynamics to defend plant viral and bacterial infection (Hatsugai et al, 2004, 2009). However, whether vacuolar acidification is involved in host–pathogen interaction is unclear.

Vacuolar H⁺‐ATPase (V‐ATPase) as the main driving force for acidification in the lumen of intracellular compartments is a conserved multiprotein complex (Cipriano et al, 2008). In eukaryotes, V‐ATPase is ubiquitous and resides within many intracellular components including vacuoles and endosomes. V‐ATPase is organized in two sections, named V₁ and V_o. Among the V‐ATPase, V₁ is a peripheral complex located on the cytosolic side of the membrane, which consists of subunits A ~ H and catalyzes ATP hydrolysis, while V_o is a membrane‐integral domain responsible for proton translocation. V₁ and V₀ are connected by peripheral stator stalks formed by VHA‐E and VHA‐G (Forgac, 2007; Schumacher & Krebs, 2010). V‐ATPase functions in processes such as endocytosis, protein processing, and degradation through endocytic or autophagic pathways (Nishi & Forgac, 2002; Cipriano et al, 2008). The functions of V‐ATPase mostly rely on its H⁺‐pumping activity; and the pH of intracellular compartments in eukaryotic cells is a strictly regulated parameter that affects many processes (Maxson & Grinstein, 2014). In plants, V‐ATPase activity acidifies both the vacuoles and trans‐Golgi network/early endosome (TGN/EE) (Luo et al, 2015). In mammal and yeast, some studies found the counteractions between pathogens and V‐ATPase‐regulated immune defense (Kissing et al, 2018). To promote pathogen infection, the acidification of the surrounding phagosome lumen was blocked by pathogens (Sturgill‐Koszycki et al, 1994; Wong et al, 2011; Levitte et al, 2016). Recently, V‐ATPase has been found to recruit ATG16L to initiate xenophagy during bacterial infection. Moreover, a bacterial effector SopF disrupts infection‐induced V‐ATPase‐ATG16L association to promote replication (Xu et al, 2019). However, it is unknown whether and how V‐ATPase is involved in plant–pathogen interaction.

Autophagy is an evolutionary conserved catabolic process for vacuole/lysosome‐mediated degradation of unwanted cellular components including dysfunctional organelles and macromolecules (Marshall & Vierstra, 2018). During autophagy, double‐membraned autophagosomes mature into vacuoles by subsequent fusion with vacuolar membranes. In the vacuoles, degradation of autophagic bodies is supposed to be triggered by abundant hydrolases in the acidified vacuolar environment, which relies on V‐ATPase function. Autophagy plays an important role in antiviral defense in plants (Yang et al, 2020). However, it is unclear whether V‐ATPase‐mediated vacuolar acidification contributes to antiviral autophagy defense.

We have demonstrated that autophagy functions as an antiviral mechanism against Barley stripe mosaic virus (BSMV) infection (Yang et al, 2018b). BSMV is a positive‐strand RNA virus consisting of three genomic segments (RNAα, β, and γ). The BSMV RNA genome encodes seven major proteins, RNAα encodes replicase αa, which contains the methyltransferase and helicase domains of the viral RNA‐dependent RNA polymerase (RdRp) complex, RNAβ encodes coat protein (CP) and triple gene block (TGB) movement proteins, RNAγ encodes replicase γa, which is the polymerase subunit of the RdRp complex, and the multifunctional protein γb (Jackson et al, 2009; Jiang et al, 2021). During BSMV infection, γb protein interacts with the autophagy‐related protein 7 (ATG7) to suppress autophagosomes formation by disrupting the interaction between ATG7 and ATG8 for effective virus infection (Yang et al, 2018b). However, it is unclear whether and how other BSMV proteins are involved in autophagy‐mediated antiviral defense.

Here, we report that BSMV replicase γa inhibits the vacuolar acidification and blocks degradation of autophagic bodies to promote viral infection by direct interacting with the vacuolar H⁺‐ATPase subunit B2 (VHA‐B2) to disrupt the interaction between VHA‐B2 and V‐ATPase catalytic subunit E (VHA‐E), thus inhibiting the V‐ATPase activity. Further our findings first link vacuolar acidification to both plant–pathogen interaction and show that vacuolar acidification is essential for plant antiviral immunity.

Results

BSMV γa directly interacts with VHA‐B2

To identify potential host BSMV γa‐interacting proteins, we transiently expressed GFP‐tagged γa (γa‐GFP) in Nicotiana benthamiana leaves, then performed immunoprecipitation (IP) using anti‐GFP antibody, followed by Q‐Exactive liquid chromatography tandem mass spectrometry (LC‐MS/MS). From this screen, VHA‐B2, a subunit of V‐ATPase that forms a core hexametric complex with subunit A to pump protons across the membrane to drive organelle acidification (Forgac, 2007; Cipriano et al, 2008; Maxson & Grinstein, 2014), which was mainly localized on tonoplast and also partially localized on ARA6‐ or ARA7‐labeled endosomes (Appendix Fig S1), was identified as a specific γa‐binding partner. We validated the VHA‐B2‐γa interaction in planta by co‐IP assays. The γa‐GFP fusion protein, but not free GFP, co‐immunoprecipitated with Myc‐VHA‐B2 (Fig 1A). To investigate whether the biological significance of γa is dependent on the γa‐VHA‐B2 interaction, we further mapped the region of γa responsible for the interaction with VHA‐B2 by co‐IP assays and found that the amino acid (aa) 505 ~ 578 region of γa was responsible for the γa‐VHA‐B2 interaction (Appendix Fig S2). Remarkably, Ala substitution at residue Arg‐569 (γa_R569A‐GFP) failed to precipitate with Myc‐VHA‐B2 (Fig 1A, Appendix Fig S3). Based on AlphaFold predicting (Jumper et al, 2021), γa showed high structural similarities to viral RNA polymerases that is consistent with previous studies (Jackson et al, 2009) and is predicted to have a C‐terminal RNA polymerase domain and an enigmatic N‐terminal extension domain. R569 located at thumb domain and on γa molecular surface, when mutating R569 to alanine reduced its affinity for VHA‐B2. We speculate R569A was defective in interface between γa and VHA‐B2 (Appendix Fig S3).

Figure 1. BSMV γa interacts with the host factor VHA‐B2.

1. Co‐IP analysis showing the interaction between γa and VHA‐B2 in vivo. N. benthamiana leaf tissues expressing proteins were harvested at 48 h post‐infiltration (hpi). Total proteins were immunoprecipitated with GFP‐trapped agarose. Input and IP were analyzed by protein gel analysis with anti‐GFP and anti‐Myc antibodies. Arrowhead indicates the specific band of the target protein.
2. Pull‐down assay showing the interaction between γa and VHA‐B2 in vitro. His‐VHA‐B2 purified from E. coli was incubated with the immunoprecipitated γa‐GFP, γa_R569A‐GFP, or GFP by GFP‐trapped agarose. The proteins were detected by protein gel blot analysis with anti‐His and anti‐GFP antibodies. Arrowhead indicates the specific band of the target protein.
3. BiFC analysis showing the interaction between γa and VHA‐B2 in vivo. nYFP‐VHA‐B2 or cYFP‐VHA‐B2 was co‐expressed with γa‐cYFP, γa_R569A‐cYFP, γa‐nYFP, or γa_R569A‐nYFP. YFP signals were visualized by confocal microscope. Bar = 20 μm.

Source data are available online for this figure.

To determine whether γa interacts directly with VHA‐B2, we performed a GFP pull‐down assay (Fig 1B). Due to unsuccessful γa expression in E. coli, we expressed γa‐GFP, γa_R569A‐GFP, or GFP, respectively, in plants by an agroinfiltration approach. At a 48‐h post‐agroinfiltration (hpi), these proteins were immunoprecipitated by GFP‐trapped agarose and further mixed with His‐VHA‐B2 that was purified from E. coli. The pull‐down assays showed that γa‐GFP, but not γa_R569A‐GFP or GFP, interacted with His‐VHA‐B2 in vitro (Fig 1B).

To figure out subcellular localization of the γa‐VHA‐B2 interaction in plants, we performed a biomolecular fluorescence complementation (BiFC) assay (Fig 1C). N‐terminal YFP (nYFP)‐tagged γa (γa‐nYFP)/γa_R569A (γa_R569A‐nYFP) and C‐terminal YFP (cYFP)‐tagged VHA‐B2 (cYFP‐VHA‐B2), or cYFP‐tagged γa (γa‐cYFP)/γa_R569A (γa_R569A‐cYFP) and nYFP‐tagged VHA‐B2 (nYFP‐VHA‐B2) were co‐expressed in N. benthamiana leaves. Western blot assays indicated the expression of all BiFC constructs (Appendix Fig S4A). Positive interaction signals (YFP fluorescence) were only observed in leaf tissues co‐expressing γa and VHA‐B2 (Fig 1C, left two panels). To further identify the precise compartment for the interaction of γa and VHA‐B2, cYFP‐VHA‐B2, and γa‐nYFP was co‐expressed with the organelle markers (Nelson et al, 2007). The confocal microscopy observations showed that the site of γa‐VHA‐B2 interaction was localized at the tonoplast but not on ARA6‐ and ARA7‐labeled endosomes (Appendix Fig S5A). Further, to test whether γa and VHA‐B2 interacted in cytosol. Myc‐GUS, γa‐Myc, and γa_R569A‐Myc was expressed in N. benthamiana leaves, respectively, and the cytosol proteins were separated as described (Chen et al, 2017). Then, co‐IP assays were performed using anti‐Myc agarose for IP and anti‐VHA‐B2 antibody for detecting endogenous VHA‐B2, the cytosol co‐IP results indicated that γa also interacted with endogenous VHA‐B2 (Appendix Fig S5B), suggesting that the location of γa‐VHA‐B2 interaction was also in cytosol.

Since barley is a natural host of BSMV, we also tested whether γa interacts with barley VHA‐B2 (HvVHA‐B2) by co‐IP assay and found that γa, but not its mutant γa_R569A and GUS control, interacted with HvVHA‐B2 (Fig EV1A).

Figure EV1. BSMV γa interacts with HvVHA‐B2 and disrupts the interaction between HvVHA‐B2 and HvVHA‐E.

1. Co‐IP analysis showing the interaction between γa and HvVHA‐B2 in vivo. N. benthamiana leaf tissues expressing proteins were harvested at 48 h post‐infiltration (hpi). Total proteins were immunoprecipitated with RFP‐trapped agarose. Input and IP were analyzed by protein gel analysis with anti‐RFP and anti‐Myc antibodies.
2. Competitive co‐IP analysis. RFP‐HvVHA‐B2 and YFP‐HvVHA‐E were co‐expressed with different amount of γa‐Myc (OD₆₀₀ = 0.5 and 1.0) or γa_R569A‐Myc (OD₆₀₀ = 1.0) in N. benthamiana leaves for 48 h. Total proteins were immunoprecipitated with GFP/YFP‐trapped agarose. Input and IP were analyzed by protein gel analysis with anti‐GFP/YFP, anti‐RFP, and anti‐Myc antibodies. Arrowhead indicates the specific band of the target protein.

Source data are available online for this figure.

Taken together, these results suggest that BSMV γa interacts with VHA‐B2 at the tonoplast and in cytosol, and the residue Arg‐569 of γa is crucial for its interaction with VHA‐B2.

Expression of BSMV γa, but not γa_R569A, disrupts the interaction between VHA‐B2 and VHA‐E

Considering that VHA‐B2 is a subunit of V‐ATPase, we investigated whether γa expression affects the interaction between VHA‐B2 and other subunits of V‐ATPase. To test this, we first screened the potential subunits of V‐ATPase binding to VHA‐B2 by LS/MS‐MS and found that V‐ATPase catalytic subunit E interacts with VHA‐B2 (Fig 2A), which is a component of connect factors between V₁ and V₀ of V‐ATPase (Schumacher & Krebs, 2010). Firstly, the subcellular localization of VHA‐B2 and VHA‐E in N. benthamiana was checked by protein fraction assay and found that VHA‐B2 and VHA‐E localized in cytosol and on the membrane (Appendix Fig S6). Since VHA‐B2 was mainly localized on tonoplast and partially on endosomes (Appendix Fig S1), we investigated which membrane VHA‐E are localized to by co‐expressing YFP‐VHA‐E with various organelle markers in N. benthamiana leaves, followed by microscopy observation. The results showed that VHA‐E localized with tonoplast, endosomes, Golgi, and plasma (Appendix Fig S7). Where the VHA‐B2‐VHA‐E interaction occurs was also investigated by co‐expressing nYFP‐VHA‐B2 and cYFP‐VHA‐E with the markers labeling ER, Golgi, plasma, endosomes, and tonoplast (Nelson et al, 2007) following by BiFC assays and found that the VHA‐B2‐VHA‐E interaction site was mainly localized on the tonoplast (Appendix Fig S8).

Figure 2. BSMV γa, but not γa_R569A, disrupts the interaction between VHA‐B2 and VHA‐E.

1. Competitive co‐IP analysis. RFP‐VHA‐B2 and YFP‐VHA‐E were co‐expressed with different amount of γa‐Myc (OD₆₀₀ = 0.5, 0.75, and 1.0) or γa_R569A‐Myc (OD₆₀₀ = 1.0) in N. benthamiana leaves for 48 h. Total proteins were immunoprecipitated with GFP‐trapped agarose. Input and IP were analyzed by protein gel analysis with anti‐GFP, anti‐RFP, and anti‐Myc antibodies.
2. BiFC analysis showing the γa protein blocks the interaction between VHA‐E and VHA‐B2. nYFP‐VHA‐B2 and cYFP‐VHA‐E was co‐expressed with different amount of γa‐Myc (OD₆₀₀ = 0.5, 0.75 or 1.0) or γa_R569A‐Myc (OD₆₀₀ = 1.0) in N. benthamiana leaves. YFP signals were visualized by confocal microscope. Bar = 30 μm.
3. Relative fluorescence intensity of BiFC analysis according to (B). The BiFC signal of nYFP‐VHA‐B2 and cYFP‐VHA‐E in γa_R569A‐Myc‐expressing plants was set as 1.0. Error bars indicate _SE from three independent experiments.

Source data are available online for this figure.

To investigate whether γa has the effect on the interaction between VHA‐B2 and VHA‐E, a competitive co‐IP assay was performed (Fig 2A). The YFP‐VHA‐E and RFP‐VHA‐B2 were co‐expressed with the γa‐Myc or γa_R569A‐Myc, respectively. After 48 hpi, total protein was extracted and purified by the GFP‐trapped agarose. The IP results showed that the interaction between VHA‐E and VHA‐B2 was gradually reduced following the gradient expression of γa protein, while the expression of the mutant γa_R569A‐Myc has no effect on the VHA‐E and VHA‐B2 interaction (Fig 2A). Further, a BiFC assay was performed to confirm this result (Fig 2B). The nYFP‐VHA‐B2 and cYFP‐VHA‐E were co‐expressed with the gradient γa or γa_R569A protein. The intensity of fluorescence of YFP was gradually reduced following the increased γa protein, while the interaction between VHA‐B2 and E was not affected by the mutant γa_R569A (Fig 2B and C). The protein expression of gradient γa and γa_R569A were confirmed by Western blot (Appendix Fig S4B). Since γa has the subcellular localization at tonoplast but not at endosomes (Appendix Fig S9) and the γa‐VHA‐B2 interaction was found at tonoplast and in cytosol, but not at ARA6‐ and ARA7‐labeled endosomes (Appendix Fig S5), γa mainly disrupts the VHA‐B2‐VHA‐E interaction at tonoplast.

In addition, we also tested the effect of γa expression on the interaction between HvVHA‐B2 and barley VHA‐E (HvVHA‐E) by a competitive co‐IP assay. RFP‐HvVHA‐B2 cannot be nonspecifically bond to GFP agarose (Fig EV1B, the first lane). More importantly, γa‐Myc, but not its mutant γa_R569A‐Myc, disturbed the interaction between RFP‐HvVHA‐B2 and YFP‐HvVHA‐E in a dose‐dependent manner (Fig EV1B).

Taken together, these data suggest that BSMV γa mainly interferes with the VHA‐B2‐VHA‐E interaction at tonoplast by competitively binding to VHA‐B2.

BSMV γa disturbs the localization of VHA‐B2 and suppresses V‐ATPase enzymatic activity

Since VHA‐B2‐containing V₁ domain is connected to membrane‐integral V₀ domain by peripheral stator stalks involved in VHA‐E (Forgac, 2007; Schumacher & Krebs, 2010) and γa protein disrupt the interaction between VHA‐B2 and VHA‐E, we investigated the influence of γa expression or BSMV infection on subcellular localization of VHA‐B2 by cell fraction and immunoblotting assays. Total proteins from the leaves expressing Myc‐GUS, γa‐Myc, γa_R569A‐Myc, or infected with BSMV were separated into soluble and microsomal membrane fractions by ultracentrifugation. Majority of endogenous VHA‐B2 protein is localized in membrane fractions in all leaf samples (Fig 3A). Nonetheless, more VHA‐B2 was found to localize in the soluble fractions in the leaves expressing γa‐Myc (approximately 19%) or infected with BSMV compared to that in the leaves expressing γa_R569A‐Myc or Myc‐GUS (approximately 6 or 7%). At 40 h post‐infection (hpi), the ratio (~ 29%) of soluble VHA‐B2 is highest (Fig 3A). Thus, γa protein partially changed the in vivo subcellular localization of VHA‐B2 from membrane to cytoplasm by its competitive interaction with VHA‐B2 against VHA‐E.

Figure 3. BSMV γa, but not γa_R569A, suppresses the V‐ATPase activity.

1. Localization of VHA‐B2 in planta. Total proteins (T) extracted from healthy, BSMV‐infected, Myc‐GUS‐expressed, or γa‐expressed plants were fractionated into soluble (S) and membrane (M) fractions by ultracentrifugation. Fractions were analyzed by protein gel blot with anti‐VHA‐B2, Myc, H⁺‐ATPase (plasma membrane marker), and Rubisco (soluble cytoplasm marker).
2. Relative V‐ATPase activity was detected in the BSMV‐infected plants and Myc‐VHA‐B2, Myc‐GUS, γa‐Myc, or γa_R569A‐Myc expressed plants. Error bars indicate SE from three independent experiments. Statistical significance determined by Student’s t‐test (*P < 0.05).
3. Western blot to detect protein expression according to (B) with anti‐Myc antibodies.
4. Expression of γa has no effect on the protein level of VHA‐B2 in vivo. VHA‐B2 level was detect by anti‐VHA‐B2 antibodies in the Myc‐GUS or γa‐Myc expression plants.
5. Representative confocal images of N. benthamiana leaves co‐expressing the vacuolar localized aleurain‐PEpHluorin and γa‐Myc, Myc‐GUS, or γa_R569A‐Myc. Bar = 20 μm.
6. Western blot assay to show the aleurain‐PEpHluorin expression of different samples from (E).

Source data are available online for this figure.

Since γa expression disrupted the interaction between VHA‐B2 and VHA‐E and changed the subcellular localization of VHA‐B2, which is a component of catalytic complex from V‐ATPase, we examined the effect of BSMV infection and γa expression on enzymatic activity of V‐ATPase. BSMV infection decreased the enzymatic activity of V‐ATPase by approx. 27% (Fig 3B, left two panels). The relative V‐ATPase activity was reduced by approximate 25% in samples expressing γa‐Myc compared to that in the sample expressing Myc‐GUS control. In sample expressing γa_R569A which failed to interact with VHA‐B2, the V‐ATPase activity was similar to that in Myc‐GUS samples (Fig 3B, right three panels). Protein expression in samples (Fig 3B) was confirmed by protein gel analysis (Fig 3C). Moreover, the expression of γa‐Myc had no effect on endogenous VHA‐B2 accumulation in plants (Fig 3D). Taken together, our data indicate that γa suppresses V‐ATPase activity by interacting with VHA‐B2 to impair the interaction of VHA‐B2 with VHA‐E and subcellular localization of VHA‐B2.

BSMV γa has no effect on endosomal pH and trafficking

As a multi‐subunit proton pump, V‐ATPase is responsible for the acidification of endosomes which play a central role in controlling the reutilization or degradation of membrane components and regulating fundamental processes including membrane turnover, immunity, development, and nutrient uptake (Scott et al, 2014). Since γa interacts with VHA‐B2 (Fig 1), we reasoned whether γa would affect endosomes. To investigate this, we first observed and counted the number of endosomes in leaves expressing Myc‐GUS or γa‐Myc, using the early and late endosomal markers ARA6‐CFP and CFP‐ARA7 (Otegui & Spitzer, 2008). We found that the number of endosomes labeled by either ARA6‐CFP or CFP‐ARA7 was similar between GUS‐ and γa‐expressing leaves (Appendix Fig S10). Next, we measured pH value in the TGN/EE and MVB/PVC by using the TGN/EE‐localized plant‐soluble‐modified ecliptic pHluorin (PEpHluorin‐BP80(Y612A)) and MVB‐localized PEpHluorin‐AtVSR2, two pH reporters which gradually loses fluorescence when pH is lowered at pH 6.2 (Shen et al, 2013). Since pH in the TGN/EE and MVB is approx. 6.2 ~ 6.3 in plants (Shen et al, 2013), the TGN and MVB dots were barely observed and had no significant differences between the N. benthamiana plant cells expressing GUS and γa (Appendix Fig S11). We further examined whether the trafficking of endosomes changed in the GUS‐ or γa‐expressing plants by observing ARA6‐CFP. No significant difference in the Myc‐GUS‐ and γa‐Myc‐expressing plant cells was found (Movies EV1 and EV2). These results, combined with the findings that γa‐VHA‐B2 interaction and γa have no endosomal localization (Appendix Figs S5 and S9), suggest that γa had no effect on endosomal pH and trafficking in plants.

BSMV γa protein disrupts plant vacuolar acidification

Since V‐ATPase can pump protons across the membrane to drive vacuole acidification to keep the activities of vacuolar hydrolases (Cipriano et al, 2008; Shen et al, 2013), we tested the effect of γa expression on vacuolar pH by using a vacuole‐localized aleurain‐PEpHluorin as a pH indicator (Humair et al, 2001; Shen et al, 2013). PEpHluorin gradually loses fluorescence as pH is lowered with fluorescence at pH 6.2, and aleurain‐PEpHluorin is not fluorescence in the normal plant vacuoles with an acidic pH ~ 5.2 in Arabidopsis (Shen et al, 2013). We found that there was barely fluorescence observed in N. benthamiana leaf tissues expressing GUS or γa_R569A (Fig 3E, right two panels), which means the pH in vacuole of N. benthamiana was lower than 6.2; however, obvious fluorescence was observed in leaf tissues expressing γa (Fig 3E, left panel). The PEpHluorin expression level was similar among three groups (Fig 3F). In addition, we also investigated whether γa expression has effect on the vacuolar acidification in barley by using aleurain‐PEpHluorin pH reporter to observe the vacuolar acidification in GUS‐, γa‐, and γa_R569A‐expressing protoplasts. Similarly, the expression of γa, but not its mutant γa_R569A or GUS, disrupted the vacuolar acidification in barley cells (Fig EV2A). The expression of these proteins was confirmed by Western blot (Fig EV2C). Further, we measured the pH value in vacuoles during BSMV or BSMV_R569A infection. For this purpose, we co‐infiltrated BSMV RNAα and RNAγ or RNAγ_R569A derivatives to initiate viral replication and observed the aleurain‐PEpHluorin expression by confocal microscope (Fig EV3). Consistent with the transient expression results shown in Fig 3E, only BSMV with wild‐type γa increased the vacuolar pH. These data suggest that the vacuolar pH in γa‐expressing cells was higher than that in the GUS control‐ or γa_R569A‐expressing cells (Fig 3E). Consistent with these findings, γa is also observed in vacuolar membrane and interacts with VHA‐B2 at tonoplast and in cytosol (Appendix Figs S5 and S9). Thus, BSMV γa protein disrupts plant vacuolar acidification by interacting with VHA‐B2 directly to affect V‐ATPase activity as the pump proton in plants.

Figure EV2. Observation of the vacuolar acidification in barley protoplast and the symptoms of BSMV and BSMV_R569A .

1. Representative confocal images of barley protoplasts transfected by the vacuolar localized aleurain‐PEpHluorin and γa‐Myc, Myc‐GUS, or γa_R569A‐Myc, respectively. Bar = 20 μm.
2. The symptoms of BSMV and BSMV_R569A infected barley at 14 dpi. Bar = 0.7 cm.
3. Western blot to show the protein expression in barley protoplasts shown in (A).
4. Western blot to show the BSMV and BSMV_R569A CP accumulation in barley. Rep indicates the repeat samples.

Figure EV3. BSMV‐infected cells have higher pH of vacuole than BSMV_R569A‐infected cells.

1. After a 24‐h expression of aleurain‐PEpHluorin, BSMV RNAα, and RNAγ or RNAγ_R569A were co‐infiltrated in the aleurain‐PEpHluorin‐expressing plants to initiate viral replication for another 24 h. Confocal analysis was conducted at 48 hpi for aleurain‐PEpHluorin expression. Bar = 15 μm.
2. Western blot to show the protein levels of aleurain‐PEpHluorin by anti‐GFP antibody.

Silencing of VHA‐B2 blocks degradation of autophagic bodies in plants

Degradation of autophagosome is supposed to depend on the vacuolar hydrolases in the acidified vacuole. To investigate the role of VHA‐B2 in autophagy, we silenced VHA‐B2 in N. benthamiana plants using the Tobacco rattle virus‐induced gene silencing (TRV‐based VIGS) approach (Liu et al, 2002). VHA‐B2‐silenced plants showed abnormal development (Fig 4A) and the vacuolar acidification is increased in the VHA‐B2‐silenced plants (Fig EV4), suggesting that V‐ATPase is a plant essential gene. Silencing of VHA‐B2 was confirmed by Western blot assays (Fig 4B, middle panel). To identify the role of VHA‐B2 in autophagy, a GFP‐ATG8 processing assay was performed to evaluate the effect of silencing of VHA‐B2 on autophagy flux (Fig 4B). When GFP‐ATG8 fusion protein is delivered to a vacuole, the ATG8 part of the fusion protein is sensitive to degradation by vacuolar hydrolases, whereas the GFP part is relatively resistant to hydrolysis, so the accumulation of free GFP can be used to monitor the autophagic flux (Shintani & Klionsky, 2004; Klionsky et al, 2021). In this study, free GFP was only detected in the TRV control plants, while in the VHA‐B2‐silenced plants, free GFP was barely detected (Fig 4B, top panel). To further confirm whether VHA‐B2‐silencing blocked the degradation of autophagosomes, we used the Cyan Fluorescent Protein (CFP)‐tagged NbATG8f (CFP‐NbATG8f) as a marker (Wang et al, 2013) to visualize autophagosomes and autophagic bodies (Fig 4C). We found that the relative level of autophagosome and autophagic bodies in VHA‐B2‐silenced plants was approx. 5‐fold higher than that in the control (Fig 4C and D). These results suggest that VHA‐B2 is required for V‐ATPase‐dependent vacuolar acidification and degradation for basal autophagic flux.

Figure 4. Silencing of VHA‐B2 reduces autophagosome and autophagic bodies degradation in plants.

1. Representative images of TRV control (left) and TRV‐VHA‐B2 (right). Bar = 7 cm.
2. GFP‐ATG8 processing assay to show autophagy level in the TRV or TRV‐VHA‐B2 samples. Pro:ATG8f::GFP‐ATG8f was expressed for 48 h. GFP‐ATG8f and free GFP were detected by anti‐GFP antibody (top panel). The protein level of VHA‐B2 was detected by anti‐VHA‐B2 antibodies (middle panel). RbcL stained by Ponceau indicates equal loading (bottom panel).
3. Representative images of autophagic activity revealed by the autophagy marker protein CFP‐ATG8f in TRV or TRV‐VHA‐B2 cells. CFP‐labeled autophagic bodies and autophagosomes are shown in cyan and autofluorescence from chloroplasts is pseudocolored in red. Bars = 10 μm. Arrowheads indicate the autophagosomes and autophagic bodies.
4. Quantification of the number of autophagosomes and autophagic bodies in TRV and TRV‐VHA‐B2 plants. The number of autophagosomes in TRV control leaves was set to 1.0. Error bars indicate _SE from three independent experiments. Statistical significance determined by Student’s t‐test (***P < 0.005).

Source data are available online for this figure.

Figure EV4. Silencing of VHA‐B2 increases pH value in plant vacuoles.

1. Representative confocal images of N. benthamiana leaves co‐expressing the vacuolar localized aleurain‐PEpHluorin and empty vector or the VHA‐B2 knock‐down vector (VHA‐B2i). Confocal analysis was conducted at 48 hpi. Bar = 15 μm.
2. Western blot to show the protein levels of aleurain‐PEpHluorin and VHA‐B2 in vivo by anti‐GFP and anti‐VHA‐B2 antibodies.

BSMV γa suppresses autophagosome and autophagic bodies degradation by interacting with VHA‐B2

Since BSMV γa disrupts plant vacuolar acidification and VHA‐B2 is involved in both vacuolar acidification and degradation of autophagic bodies, we tested the effect of γa on degradation of autophagic bodies by observing the CFP‐ATG8f‐labeled autophagosomes and autophagic bodies in Myc‐GUS, γa‐Myc or γa_R569A‐expressing leaf tissues. The number of autophagosomes and autophagic bodies was similar in γa‐Myc expressing cells between with and without protease inhibitor E‐64d treatments which inhibits vacuole‐mediated proteolysis to block the degradation of autophagic bodies (Klionsky et al, 2021), but approximately six times higher than that in Myc‐GUS expressing cells with E‐64d treatment (Fig 5A and B). However, the number of autophagosomes and autophagic bodies was similar in Myc‐GUS and γa_R569A‐Myc expressing plants either with or without E‐64d treatments, but ~ 1.5 times higher in those leaves with E‐64d treatment than that in those leaves without E‐64d treatment (Appendix Fig S12A and B). Furthermore, the CFP‐ATG8f‐labeled autophagic bodies in the γa‐expressing plants were localized in vacuoles (Movie EV3 and 4).

Figure 5. BSMV γa protein induces the aggregation of autophagic bodies in vacuole and inhibits autophagy flux.

1. Representative images of autophagic activity revealed by the autophagy marker protein CFP‐ATG8f in cells transiently expressing Myc‐GUS or γa‐Myc with or without the hydrolase inhibitor E‐64d treatment. CFP‐labeled autophagic bodies and autophagosomes are shown in cyan and autofluorescence from chloroplasts are pseudocolored in red. Bars = 10 μm. Arrowheads indicate the autophagosomes and autophagic bodies.
2. Quantification of the number of autophagosomes and autophagic bodies shown in (A). The number of autophagosomes in Myc‐GUS expression leaves was set to 1.0. More than 200 mesophyll cells were analyzed per treatment. Error bars indicate SE from three independent experiments. Statistical significance determined by Student’s t‐test (***P < 0.001).
3. Analysis of autophagic flux probed with anti‐NBR1 antibodies in Myc‐GUS‐ and γa‐Myc‐expressing plant leaves (top panel). The expression of Myc‐GUS and γa‐Myc were detected by anti‐Myc antibodies (middle panel). Large subunit of RuBisCO (RbcL) stained by Ponceau indicates equal loading (bottom panel).
4. GFP‐ATG8 processing assay to show the autophagy level in the Myc‐GUS or γa‐Myc expression samples. GFP‐ATG8f was co‐expressed with Myc‐GUS or γa‐Myc for 48 h. GFP‐ATG8f and free GFP were detected by anti‐GFP antibody (top panel). The expression of Myc‐GUS and γa‐Myc were detected by anti‐Myc antibodies (middle panel). RbcL stained by Ponceau indicates equal loading (bottom panel).

Source data are available online for this figure.

To further test the effect of γa protein on autophagy, we determined autophagy levels in γa‐Myc‐ or Myc‐GUS‐expressing plants by detecting the protein level of selective autophagy cargo receptor NBR1/Joka2, which has been established to be a selective autophagy substrate and a suitable autophagy marker for autophagic flux detection in plants (Zhou et al, 2013; Xu et al, 2017). In the γa‐Myc‐expressing leaves, the protein level of NBR1 dramatically increased compared to that in Myc‐GUS expressing leaves (Fig 5C), suggesting that the autophagy flux in γa‐expressing plants was blocked. This was also confirmed by GFP‐ATG8 processing assay (Fig 5D). The free GFP was only detected in the Myc‐GUS‐expressing plants, while in the γa‐Myc‐expressing plants, the free GFP was hardly detected (Fig 5D). These results suggest that BSMV γa protein inhibits autophagy flux in plants. Thus, BSMV γa protein suppresses autophagy by disturbing the degradation of autophagosomes and autophagic bodies.

Next, we analyzed the number of autophagosomes labeled by CFP‐ATG8f in γa‐ or γa_R569A‐expressing plants. As expected, the number of autophagosome in γa‐expressing leaf tissues was about 4‐fold higher than that in leaf tissues expressing GUS control (Fig 6A and C). However, the number of autophagosomes was similar between γa_R569A‐ and GUS‐expressing leaf tissues (Fig 6A and C). Further, NBR1 protein level was similar between γa_R569A‐ and GUS‐expressing leaf tissues, but twice less than that in γa‐expressing leaf tissues (Fig 6B). These data suggest that BSMV γa suppresses the degradation of autophagosomes by interacting with VHA‐B2.

Figure 6. Mutant protein γa_R569A failed to block the degradation of autophagic bodies and to reduce the autophagic flux.

1. Representative images of autophagic activity revealed by the autophagy marker protein CFP‐ATG8f in cells transiently expressing Myc‐GUS, γa‐Myc, and γa_R569A‐Myc. CFP‐labeled autophagic bodies and autophagosomes are shown in cyan and autofluorescence from chloroplasts are pseudocolored in red. Bars = 10 μm. Arrowheads indicate the autophagosomes and autophagic bodies.
2. Analysis of autophagic flux probed with anti‐NBR1 antibodies in Myc‐GUS‐, γa‐Myc‐, and γa_R569A‐Myc‐expressing leaves (top panel). The expression of Myc‐GUS, γa‐Myc, and γa_R569A‐Myc were detected by anti‐Myc antibodies (middle panel). Large subunit of RuBisCO (RbcL) stained by Ponceau indicates equal loading (bottom panel).
3. Quantification of the number of autophagosomes and autophagic bodies in Myc‐GUS‐, γa‐Myc‐, and γa_R569A‐Myc‐expressing plants. The number of autophagosomes in Myc‐GUS‐expressing leaves was set to 1.0. Error bars indicate SE from three independent experiments. Statistical significance determined by Student’s t‐test (**P < 0.01, ***P < 0.001).

We also assessed whether overexpression of VHA‐B2 alleviates γa‐mediated autophagy inhibition. For this purpose, CFP‐ATG8f was used to monitor the autophagosomes and autophagic bodies in leaf tissues co‐expressing γa‐Myc with Myc‐VHA‐B2 or Myc‐GUS (Fig 7A). We found that relative autophagosome number in leaf tissues co‐expressing Myc‐VHA‐B2 and γa‐Myc was much lower than that in leaf tissues agroinfiltrated with γa and empty vector, but was similar to that in leaf tissues expressing GUS control (Fig 7B). These data indicate that overexpression of VHA‐B2 recovers autophagy level in γa‐expressing leaf tissues.

Figure 7. Expression of VHA‐B2 recovers the level of autophagosomes in γa‐expressing plants.

1. Confocal microscope to observe the CFP‐labeled autophagosomes or autophagic bodies in the Myc‐GUS, γa‐Myc, or γa‐Myc and Myc‐VHA‐B2 co‐expression plant leaves, respectively. Bar = 20 μm. Arrowheads indicate the autophagosomes and autophagic bodies.
2. Quantification of the number of autophagosomes and autophagic bodies shown in (A). The number of autophagosomes in Myc‐GUS expression leaves was set to 1.0. More than 200 mesophyll cells were analyzed per treatment. Error bars indicate SE from three independent experiments. Statistical significance determined by Student’s t‐test (***P < 0.001).

Taken together, our data suggest that BSMV γa suppresses the degradation of autophagic bodies by blocking vacuolar acidification through its interaction with VHA‐B2 to inhibit V‐ATPase activity in plants.

The BSMV γa‐VHA‐B2 interaction contributes to effective viral infection

To explore the biological significance of the γa‐VHA‐B2 interaction in plant antiviral defense against BSMV, BSMV_R569A mutant virus was generated. When N. benthamiana plants were infected, BSMV_R569A and wild‐type BSMV caused a similar symptom in young systemic leaves at early stage of infection (10 days post‐inoculation, dpi) (Fig 8A). However, at the late infection stage (30 dpi), newly emerging systemic young leaves of N. benthamiana plants infected with BSMV_R569A lost symptoms and accumulated no detectable level of viral coat protein (CP) (Fig 8B and C). In the early stage of infection (10 dpi), less CP accumulation were detected in inoculated leaves and young systemic leaves infected with BSMV_R569A than that of wild‐type BSMV‐infected plants (Fig 8C). Co‐infection with RNAα and RNAγ is sufficient for BSMV replication but without viral movement (Zhang et al, 2017). We measured BSMV negative‐strand RNA level in the inoculated leaves agroinfiltrated with RNAα and RNAγ or RNAγ_R569A by real‐time RT–PCR and found less viral RNA accumulation in BSMV_R569A mutant virus‐infected plants than wild‐type BSMV‐infected plants (Fig 8D), suggesting that γa‐VHA‐B2 interaction probably contributes to effective BSMV replication. To investigate whether the effect of R569A mutation in γa on BSMV accumulation is due to the inability to inhibit vacuolar acidification, we measured BSMV negative‐strand RNA level in the inoculated VHA‐B2‐silenced N. benthamiana leaves infected with RNAα and RNAγ or RNAγ_R569A by real‐time RT–PCR and found similar BSMV RNA levels in the VHA‐B2‐silenced leaves infected with BSMV and BSMV_R569A (Fig 8E), suggesting that R569A mutation in γa has no effect on BSMV replication ability in VHA‐B2‐silenced plants. Thus, the less BSMV_R569A accumulation in wild‐type plants is due to the inability to inhibit vacuolar acidification. We confirmed the effect of γa‐VHA‐B2 interaction on BSMV infection in barley plants (Fig EV2B and D). BSMV_R569A caused milder viral symptom than wild‐type BSMV in barley plants, and viral protein accumulation in barley plants infected with wild‐type BSMV was higher than that in barley plants infected with BSMV_R569A (Fig EV2B and D). These results suggest that the γa‐VHA‐B2 interaction contributes to effective BSMV infection and functional vacuolar acidification is required for plant antiviral defense against BSMV.

Figure 8. Mutant virus BSMV_R569A causes reduced infection in plants.

1. Systemic symptoms of BSMV or BSMV_R569A at 10 dpi. Bar = 7 cm.
2. Systemic symptoms of BSMV or BSMV_R569A at 30 dpi. Bar = 7 cm.
3. The CP proteins of BSMV or BSMV_R569A extracted from infiltrated leaves at 3 dpi and from systemic leaves at 10 dpi or 30 dpi were detected by Western blot using anti‐CP antibodies, respectively. The mutant virus BSMV_R569A has less viral CP accumulation in the infiltrated leaves at 3 dpi, and in the upper systemic leaves at 10 dpi and 30 dpi.
4. Real‐time RT–PCR analysis of BSMV gRNA levels in the TRV control infiltrated with BSMV RNAα and RNAγ or RNAγ_R569A. The negative‐stranded RNAγ was used to indicate BSMV gRNA level, the house‐keeping gene PP2A served as the internal control. The data are presented as the relative ratio of gene expression compared with RNAα + RNAγ, which was set to 1.0. Error bars indicate SE from three independent experiments. Statistical significance determined by Student’s t‐test (*P < 0.05).
5. Real‐time RT–PCR analysis of BSMV gRNA levels in the VHA‐B2‐silenced leaves infiltrated with BSMV RNAα and RNAγ or RNAγ_R569A. The negative‐stranded RNA of γ was used to indicate BSMV gRNA level, the house‐keeping gene PP2A served as the internal control. The data are presented as the relative ratio of gene expression compared with RNAα + RNAγ, which was set to 1.0. Error bars indicate SE from three independent experiments.

Vacuolar acidification changes during the infection with diverse plant viruses

To investigate whether subversion of vacuolar acidification is a general counter‐defense strategy during plant virus infection, the aleurain‐PEpHluorin expressed in plants infected with BSMV, Lychnis ringspot virus (LRSV), Cucumber mosaic virus (CMV), or Potato virus X (PVX) was observed by confocal macroscope. We found higher vacuolar pH in the cells infected with BSMV, LRSV, and CMV, but not with PVX, compared to the control cells without virus infection (Fig EV5). These data suggest that multiple viral infections disrupt vacuolar acidification to counter antiviral defense in plants.

Figure EV5. Vacuolar acidification changes during the infection with multiple plant viruses.

1. Representative confocal images of N. benthamiana leaves co‐expressing the vacuolar localized aleurain‐PEpHluorin with empty vector, CMV, PVX, and LRSV. Bar = 20 μm.
2. Western blot to show the expression of aleurain‐PEpHluorin by anti‐GFP antibody.

Source data are available online for this figure.

Discussion

In this study, we showed that a plant RNA virus encodes a protein to suppress vacuolar acidification and block the degradation of autophagosomes, to establish an effective infection by impairing the interaction between VHA‐B2 and VHA‐E and subcellular localization of VHA‐B2 through its binding to VHA‐B2 to disrupt the V‐ATPase activity. To our knowledge, this is the first report that vacuolar acidification is essential for plant defense and that a plant–pathogen interferes with vacuolar acidification and counters antimicrobial autophagy defense to promote its infection by blocking vacuolar acidification and the degradation of autophagic bodies in plants.

V‐ATPases are ubiquitous multi‐subunit complexes mediating the acidification of many intracellular compartments including vacuoles and endosomes in eukaryotic cells (Kane, 2006, 2012; Forgac, 2007). In mammals and humans, various pathogenic agents including enveloped viruses like influenza and Ebola viruses, take advantage of the acidic endosomes to activate their ability to penetrate cell membrane and entry into host cells (Kielian & Jungerwirth, 1990; Cotter et al, 2015). The V‐ATPases‐mediated acidification of endosomes is essential for a vast majority of human viruses such as human coronavirus NL63, influenza viruses, Zika virus, dengue virus, Sindbis virus for their successful entry and release of the viral genome into the cytoplasm (Hunt et al, 2011; Jang et al, 2018; Kao et al, 2018; Milewska et al, 2018; Owczarek et al, 2019). However, influence of acidification of vacuoles or endosomes on plant–pathogen invasion has never been studied. Our data showed, for the first time, that BSMV infection and its replicase γa increased the vacuolar pH value in plants (Figs 3 and EV3); thus, the enzymatic activities of hydrolases in vacuoles were inhibited, which led to the autophagic bodies accumulated in the vacuoles (Fig 5). This is consistent with our observation that γa, but not its mutant γa_R569A, partially changes the subcellular localization of VHA‐B2 from membrane to cytoplasm (Fig 3) by competitively disrupting the interaction between VHA‐B2 and VHA‐E (Fig 2) and co‐localized with the vacuolar membrane (Appendix Fig S9), thus interfering with the proper assembly of functional V‐ATPase at vacuolar membrane. This is also consistent with our observation that the VHA‐E‐VHA‐B2 interaction mainly occurred on tonoplast (Appendix Fig S8) and the γa‐VHA‐B2 interaction was found at tonoplast (Appendix Figs S5A and S9) and into cytosol (Fig 3A, Appendix Fig S5B). Since two V‐ATPase domains V₁ and V₀ are connected by the peripheral stator stalks formed by VHA‐E and VHA‐G (Schumacher & Krebs, 2010) and the formation of V‐ATPase is a reversible process as a enzymatic activity regulatory mechanism (Kane, 2006, 2012; Forgac, 2007), BSMV γa could impair the connection between V₁ and V₀ domains, thus disrupt the formation of V‐ATPase on tonoplast by competitively binding to the VHA‐B2 to detach VHA‐B2 from tonoplast into cytosol. Further, the mutant virus BSMV_R569A, which contains a γa mutant that fails to interact with VHA‐B2, possesses less viral accumulation, and causes milder viral symptom (Figs 8 and EV2B and D, Appendix Fig S3), suggesting that vacuolar acidification is essential for plant defense. In addition, we found that the infection of multiple viruses increased the vacuolar pH value in plants (Fig EV5), suggesting that vacuolar acidification could contributes to plant defense against multiple viruses. In plants, V‐ATPase activity is also essential to acidify the TGN/EE (Luo et al, 2015). However, we did not observe that γa affected the pH value, trafficking, and the number of endosomes in plants (Appendix Figs S10 and S11, and Movies EV1 and EV2). These findings are consistent with the observations that the sites of γa‐VHA‐B2 interaction do not co‐localize with endosomes (Appendix Fig S5A) and that γa has been reported to be localized around the chloroplasts (Zhang et al, 2017) and the vacuolar membrane (Appendix Fig S9). Vacuolar acidification is a key process for host cells for maintaining diverse physiological functions such as degrading dysfunctional organelles or cytoplasmic components. Consistent to this concept, silencing of VHA‐B2 caused abnormal plant developmental phenotype (Fig 4A). Disruption of vacuolar acidification could directly suppress autophagy‐dependent antiviral defense to promote virus infection. Indeed, disruption of vacuolar acidification blocks autophagy (Fig 4). There are some other additional possibilities, for example, disruption of vacuolar acidification may affect the normal functions of host factors involved in plant antiviral defense or vacuolar degradation of viral RNAs or proteins, thus promotes virus infection.

V‐ATPase has been reported to regulate antibacterial defense in mammal (Xu et al, 2019). During autophagy, the double‐membraned autophagosomes mature and fuse with lysosomal or vacuolar membranes. In the autolysosome, the degradation of components in autophagosomes is triggered by the hydrolases in the acidified environment which is regulated by V‐ATPase activity (Forgac, 2007; Marshansky & Futai, 2008; Kissing et al, 2018). For example, when Salmonella infection causes vacuolar damage, host V‐ATPase recruits ATG16L1 to initiate antibacterial autophagy. However, Salmonella effector SopF disrupts V‐ATPase‐ATG16L1 association to promote bacterial replication (Xu et al, 2019). This finding reveals a direct connection between V‐ATPase and antibacterial autophagy defense. In this study, we first reveal that a viral effector interacts with and inhibit V‐ATPase activity to block vacuolar acidification for the effective viral infection. Further, multiple viruses are found to disrupt vacuolar acidification (Fig EV3), suggesting that the disruption of vacuolar acidification is a general mechanism for viruses to counter plant antiviral defense.

Autophagy plays an important role during plant virus infection (Ismayil et al, 2020b; Yang et al, 2020). Apart from its antiviral defense (Hafren et al, 2017; Haxim et al, 2017; Hafrén et al, 2018; Li et al, 2018, 2020), pathogens have evolved various mechanisms to evade or degrade host defense‐related proteins to promote infection (Baumberger et al, 2007; Fu et al, 2018; Li et al, 2019; Michaeli et al, 2019). Virus‐triggered autophagy can promote plant fitness and reduce symptom severity during DNA and RNA virus infections. Autophagy‐mediated plant fitness is also necessary for a long‐term virus survival (Hafren et al, 2017; Hafrén et al, 2018; Ismayil et al, 2020a). We previously demonstrated that BSMV γb inhibits autophagy by directly interacting with ATG7 to disrupt the ATG7‐ATG8 complex (Yang et al, 2018b). In this study, we found another new mechanism of virus‐mediated suppression of autophagy, by which a viral protein inhibits the autophagic bodies degradation by blocking vacuolar acidification via interfering with V‐ATPase activity. It needs further investigation whether other pathogens use similar mechanism to inhibit the vacuolar acidification for their infection.

Materials and Methods

Plant growth

Nicotiana benthamiana plants were grown in a climate‐controlled chamber at 24°C with a 16/8‐h light/dark photoperiod with light intensity of approx. 75 mmol/m²s provided by three cool daylight tubes.

Plasmid construction and virus infection

Barley stripe mosaic virus γa gene and its mutant were cloned into T‐DNA vector pLIC‐Myc (Ismayil et al, 2020a), respectively, to generate γa‐Myc and γa_R569A‐Myc. Full‐length NbVHA‐B2 was amplified and cloned into pMyc‐LIC vector to generate Myc‐VHA‐B2. The γa‐GFP, γa‐nYFP, and γa‐cYFP constructs have been described (Zhang et al, 2017). The full‐length VHA‐B2 and the C‐terminal γa (γa_F3) were cloned into pET28a to express His‐tagged fusion protein in E. coli. Full‐length VHA‐B2 was cloned into pLIC‐NE/CE or pNE/CE‐LIC vectors, respectively, to generate the BiFC‐related vectors. The VHA‐B2‐hairpin RNA construct for VHA‐B2‐silencing in plants was constructed as (Xu et al, 2010). Virus infection was performed as described (Zhang et al, 2017; Jiang et al, 2018). The primers used for these vectors are shown in Table S1.

Co‐IP

Co‐IP assays were performed as described (Zhang et al, 2017). Expression vectors were co‐agroinfiltrated into N. benthamiana leaves. The infiltrated leaves were harvested at 48 h post‐infiltration (hpi), and total proteins from 2 g of leaf tissues were ground in a liquid nitrogen‐cooled mortar and mixed with the extraction buffer (10% [v/v] glycerol, 25 mM Tris–HCl, pH 7.5, 1 mM EDTA, 150 mM NaCl, 0.2% NP40, and protease inhibitor cocktail). The supernatant was incubated with GFP beads (ChromoTek) for 4 h at 4°C. The precipitates were washed three times and analyzed by immunoblotting using anti‐GFP (ChromoTek) and anti‐Myc (Abmart) antibodies.

Pull‐down assays

GST pull‐down assays were performed as described (Yang et al, 2018a). For pull‐down assay between γa‐GFP and VHA‐B2‐His, the γa‐GFP infiltrated leaves were harvested at 48 hpi, 3 g of leaf tissues were ground in a liquid nitrogen‐cooled mortar and suspended with extraction buffer, the supernatant was incubated with GFP beads. The GFP bead‐bound γa‐GFP was co‐incubated with purified His‐VHA‐B2 protein in vitro in extraction buffer for 3 h at 4°C. The precipitates were analyzed by immunoblotting using anti‐GFP and anti‐His (Abmart) antibodies.

Protein fraction assay and quantification

A plasma membrane protein isolation kit (Invent Biotechnologies, SM‐005‐P) was used for the protein fraction analysis according to the manufacture’s protocol. Briefly, the fresh plant leaves (~ 250 mg), 100 μl Kit buffer A, and 60 mg tissue dissociation bead were added into the filter cartridge for grinding. After grinding, Kit buffer A was added to the filter to the top, followed by keeping the filters on ice for 5 min. After centrifuge at 16,000 g for 30 s, discard the filter. After vortexing and then another centrifuge at 700 g for 1 min, partial supernatant was taken as the total protein (T). Transfer the remaining supernatant into a fresh 1.5‐ml microcentrifuge tube and centrifuged at 16,000 g for 45 min at 4°C, the resulting supernatant (S) is the cytosolic fraction, and the pellet (M) is the total membrane fraction. The S part was transferred to a new 1.5‐ml microcentrifuge tube and has the same diluted ratio with total protein. The pellet was washed by the buffer A for five times to remove the potential supernatant pollution and then was diluted by buffer A with the same diluted ratio with total protein. Finally, the proteins were analyzed by Western blot assay and ImageJ. The total protein (T) was set as 100%, and the cytosolic fraction (S) and the membrane proteins (M) was calculated by ImageJ software.

V‐ATPase activity assay

V‐ATPase activity assay was conducted as previously described (Heinonen & Lahti, 1981; Krebs et al, 2010; Tang et al, 2012). Plant leaves were ground with cold homogenization buffer containing 350 mM sucrose, 70 mM Tris–HCl (pH 8.0), 3 mM Na₂EDTA, 0.2% (w/v) BSA, 1.5% (w/v) PVP‐40, 5 mM DTT, 10% (v/v) glycerol, and 1× protease cocktail inhibitor mixture (Roche). The homogenate was filtered through four layers of cheesecloth and centrifuged at 4,000 g for 20 min at 4°C. The supernatant was filtered through cheesecloth again and then centrifuged at 100,000 g for 1 h. V‐ATPase activity of 10 μg microsomal membranes was determined as phosphate (Pi) release after 40 min incubation at 28°C. The resulting microsomal pellet was resuspended in 350 mM sucrose, 10 mM Tris‐MES (pH 7.0), 2 mM DTT, and 1× protease inhibitor mixture. The V‐ATPase assay solution contained 25 mM Tris‐MES (pH 7.0), 4 mM MgSO₄, 50 mM KCl, 1 mM NaN₃, 200 μM Na₂MoO₄, 500 μM NaVO₄, 0.1% Brij 35, and 2 mM Mg‐ATP. For measuring the amount of inorganic Pi, reactions were terminated by adding 40 mM citric acid. Freshly prepared AAM solution (50% (v/v) acetone, 2.5 mM ammonium molybdate, 1.25 M H₂SO₄) was then added to the reaction, vortexed, and colorimetrically examined at 355 nm.

Confocal observation of autophagosomes, autophagic bodies, and pH detection

The related constructs and the autophagy marker CFP‐ATG8f were co‐infiltrated into N. benthamiana leaves. At 48‐h post‐infiltration, 20 mM protease inhibitor E‐64d was infiltrated in the same area. 8 h later, the samples were observed and imaged using Zeiss 710 microscope (Wang et al, 2013). The number of autophagic bodies and autophagosomes were counted by ImageJ. pH detection by observation of aleurain‐PEpHluorin was performed as described (Tamura et al, 2003; Martiniere et al, 2013). The aleurain‐PEpHluorin and Myc‐GUS, γa‐Myc, or γa_R569A‐Myc were co‐agroinfiltrated in the N. benthamiana leaves in the three area of the same leaf or transfected into barley protoplasts, respectively. The images were captured after a 48‐h expression by Leica TCS SP8 STED confocal microscope with an excitation wavelength of 488 nm, and the emission was captured at 512 nm (Shen et al, 2013).

BiFC and detection of the expression of BiFC constructs

BiFC assay was performed using Leica TCS SP8 STED confocal microscope; YFP was excited at 546 nm. Images were analyzed with Imaris. The different groups containing the experimental and control group were infiltrated in the same leaf to reduce differences of expression conditions. The expression of BiFC constructs was detected by Western blot with anti‐GFP antibody (Chromotek, RRID: AB_2749857).

Real‐time PCR analysis

cDNA was synthesized from 2 μg total RNA (DNase‐treated) using an oligo(dT) and reverse transcriptase (Transgen). The gene fragments were amplified using 2× SsoFast Eva‐Green Supermix (Bio‐Rad) and the primers are shown in Table S1. PP2A was used as the internal control (Liu et al, 2012), and the data were analyzed using CFX Manger (Bio‐Rad).

Preparation and transfection of barley mesophyll protoplasts

Barley mesophyll protoplasts were prepared as description (Ohsato et al, 2003). Seedlings of barley were grown at 25°C with 16 h of light for 6 ~ 7 days. The second leaf was cut at about 5 cm from the ground surface and the epidermis was peeled off carefully by hand. Exposed mesophyll tissues were immediately placed in enzyme solution including of 0.65 M mannitol (Sigma‐Aldrich), 2% Cellulase Onozuka R‐10 (Yakult), and 0.1% BSA, pH 5.7 for a 3‐h incubation at 30°C in dark. The enzyme solution containing released protoplasts was collected and centrifuged at 100 g for 5 min. The pellet was suspended in 5 ml 0.65 M mannitol, which was layered on 20% sucrose in a 15‐ml tube, followed by centrifugation at 100 g for 8 min. Protoplasts at the interface were collected, suspended, and washed twice in 0.65 M mannitol by centrifugation at 100 g for 5 min. The final pellet was suspended in 0.65 M mannitol at a concentration of 4.5 × 10⁵ cells per ml. 1 ml protoplast suspension (containing 4.5 × 10⁵ cells) was centrifuged at 100 g for 5 min and the supernatant was removed. For protoplast transfection, the pelleted cells were suspended in 20 μl 0.65 M mannitol and 5 μg of plasmids were added, followed by immediate mixing with 120 μl PEG solution (40% polyethylene glycol, average Mr 1450 from Sigma‐Aldrich, 30 mM CaCl₂, pH 5.5). The mixture was placed on ice for 40 s, followed by addition of 1 ml 0.65 M mannitol and incubation on ice for 30 min. Inoculated protoplasts were washed twice in 0.65 M mannitol by centrifugation and suspended in 1 ml of medium consisting of 0.2 mM KH₂PO₄, 1 mM KNO₃, 1 mM MgSO₄, 1 mM KI, 0.1 mM CuSO₄, 10 mM CaCl₂, 0.65 M mannitol, pH 6.5. The cell suspension was transferred to a 24‐well plate (Corning) and incubated at 17°C for up to 48 h in the dark.

Statistical analysis

Data shown in this study are means of three independent experiments and analyzed by Student’s t‐test. Protein band signals were quantified by using ImageJ software. Quantifications for the localization/colocalization photographs were calculated by the Leica TCS SP8 STED confocal microscope.

Accession numbers

Sequence data from this article can be found at Sol Genomics Network (https://solgenomics.net) and Nicotiana benthamiana Genome & Transcriptome database (http://benthgenome.qut.edu.au/) in the under the following accession numbers: NbVHA‐E (Niben101Scf04673g01003.1), NbVHA‐B2 (Niben101Scf01718g05001.1), HvVHA‐B2 (L11873.1).

Author contributions

YL, MY, and DL conceived the project. MY, AI, YW, XZ, and ZJ designed the experiments, which mainly performed by MY. LY predicted the structure of γa and its mutant protein. All authors analyzed the data. MY, YL, YH and DL wrote the article.

Supporting information

Appendix

Expanded View Figures PDF

Movie EV1

Movie EV2

Movie EV3

Movie EV4

Source Data for Expanded View/Appendix

Source Data for Figure 1

Source Data for Figure 2

Source Data for Figure 3

Source Data for Figure 4

Source Data for Figure 5

The EMBO Journal (2022) 41: e108713

Data availability

This study includes no newly generated data deposited in external repositories.